Go with the Data Flow: Maximizing Tunnel Operation and Maintenance

BART DUIJVESTIJN: Hi to you all. Go with the Data Flow Maximizing Operation and Maintenance. That's the question we did, and we are aware that we are on a high speed train for digital development. But are we driving in fog? We do not think so.

We think that it's possible to hop on a train and just think to maximize operation and maintenance. How do we do it? We did making it a digital twin from all the data which we use for maintenance, and that's interesting. Some years ago, we had a discussion with Tunneling International.

It was a newspaper, and it was asking us Digital Twins, Hope or Hype? Do you know in your practice what digital twins are and what it is? Please see it for yourself, what is it? And then when we doing some things on digital twins, are they all the same or is a digital twin something which differs from the questions there are and is it all clear or is it some uncertainty within what is a digital twin?

We have some solutions, and we did a way on working it and we'd like to share it with you. And that's the reason why we are telling something about the digital twin we do for tunnel management. Some years ago, we started with it and a client asked about the digital twin. What is it? I don't need a digital picture of my tunnels.

If I want to see my tunnels, I go to them and see them in real. So I do not need the digital twin, but we made another thing using digital twins. Interest and excitement, the lack of definitions of a digital twin is, but what it is work more efficiently, more effective, and to prevent unnecessary errors, saving time and money. It was possible to do it.

We'd like to take you with us and show what we made for digital twin. Please, let's go on and hop on the train with us. My name is Bart Duijvestijn. I'm working as head of the Department Tunnels of Arcadis, a global company who works all over the world. And I have a [INAUDIBLE] in probabilistic design. I did since my study and like to make decisions in the uncertainties, and the digital twin will help with it, and we talk about it together, Jurjen and I, and now I would like to give the talk to Jurjen.

JURJEN HAITSMA: Good afternoon. My name is Jurjen Haitsma. I'm from the Netherlands, and I am a project manager within Arcadis. I was 22 years in the construction industry and last five years I'm program manager for digital in Arcadis. Together with Bart, we experienced the journey of the digital twin of the Waterwolftunnel. In this presentation, we will introduce our clients, we will introduce how we optimize our operation and maintenance of the tunnel process, and Bart is going to do a deep insight into use cases, tunnel lightning, and tunnel ventilation.

We will finalize this presentation with a few on the future. How are we going to proceed? But first, something about Arcadis. Arcadis is a global design and engineering and consultancy company active in 70 countries. We collaborate to improve quality of life, and we want to maximize our impact in our business areas' resilience, places, mobility, and intelligence, our digital product workstream.

We are not alone. We are with more than 36,000 people. We work in 70 countries, and this is how we want to realize our strategy.

[VIDEO PLAYBACK]

[MUSIC PLAYING]

- We continue to see significant changes to the world around us. Changes to our urban landscape, climate change putting pressure on resources and requiring cities and businesses to become more resilient, demands on housing, commercial property, and infrastructure, and the potential to empower data, unlocking decision making and automation for our clients. These trends continue to accelerate, increasing clients demands and shaping our future strategy.

To maximize that impact, Arcadis are focusing on three key business areas. Helping clients to use, renew, and protect essential resources, capitalizing on our experience in North America. Through human-centered design, we will create the spaces people need to work and live, building on our successes in Asia-Pacific, and keeping cities and people moving and connected, as we are doing across Europe. These complex challenges require the proven expertise of Arcadis.

Over the next three years, we will maximize our impact by putting sustainable solutions at the heart of what we do, leaving clients more resilient, effective, and secure. By simplifying operations and focusing on leveraging our experience and global scale, and we will continue our drive towards digital leadership, targeting investment on digital products and services and investing in the skills of Arcadis to maximize their impact. In three years' time, we will be measuring the impact of all our projects against the United Nations Sustainable Development Goals.

Our clients will know us for giving them access to digital solutions that maximize their impact. And we will have a reputation for the best training when it comes to skills in digital and sustainable leadership. We are Arcadis. We exist to improve quality of life.

[MUSIC PLAYING]

[END PLAYBACK]

JURJEN HAITSMA: And that's what Bart and I also stands for. This is what we as Arcadis do for our asset owners. That's where we stand up from our beds every day, and this project is spot on the summary of our strategy. From a digital perspective, we hit all the boxes.

Unlocking data for clients, facilitate data-driven decision making, develop and deliver digital service and solutions. In this way, we achieve our mission, improving quality of life. And data is a key factor particular in operate and main processes. Let's see how this comes into practice.

In our tunnel project in the Netherlands. These are the key items of our project. The asset owner is the province of North Holland. They own the three tunnels, and as Arcadis, we are the managing agent. That means we are strategic director from complex critical infrastructure.

And we can do the job because we have specific knowledge about it, and we do it in a cost effective way. We do it since 2013. We do the management of three objects, two tunnels, and a special, an aqueduct. They were designed and built in different periods, so there are different specification and installation technology.

All the objects are critical to the traffic network around Schiphol, our International Airport and Amsterdam, the capital of the Netherlands. You see how many traffic passes the tunnel every year. The Waterwolftunnel, the biggest amount, more than 30 million traffic units.

This is the way how we organize this collaboration with our clients. The client is responsible for the strategic level. Arcadis is moving strategic into tactics and translate it together with our partners in planning and work preparation. The tool of our collaboration is to provide reliable, available, and safe mobility through the tunnels.

All activities we explore are designed to provide flow and avoid disruptions. Right data at the right time is crucial for an effective operation, and we'd like to see this collaboration process from the point of view of the digital asset lifecycle. This is our language to connect processes from the start of the project, figure design construction operation maintenance to retrofit of an asset.

In the six different phases, we have specialized products and services. Let me make it explicit for three phases. Data-driven design, data-driven construction, and data-driven operation. In the left boxes you see the products and services that we provide. In the right boxes, we see data reuse, and that's the core of our approach to store data one time and reuse that data every phase as much as possible.

To create that for our tunnel, this is the process we went through. We reach the Netherlands, we move into Amsterdam, and Schiphol. The start of our data model is a point cloud. The point cloud to a data BIM model was realized by automatic object recognition, and doing so, we created a Revit model, a model we can use for visualization and simulation.

You'll see all the installations, the lights, the ventilation well positioned on the right place in the tunnel. The installations can also be used for safety training and training with the fire brigade, the police, and the ambulance. The geometric model we also use to identify critical data points, which we connect with static and dynamic data.

To create a digital asset lifecycle, we see five steps. The first step to define valuable data points, the second step to gather them with critical data, and then visualize and make control limits. The fourth step is to evaluate the detail and the frequency, and the last but not least step is to communicate up and downstream the process the effects of the measurements.

And as you see, we have people below. We do not do that alone. We work intensively with our clients and our partners to make this work in a valuable way, and how that becomes practice will be presented by Bart. I will now hand over to Bart.

BART DUIJVESTIJN: So to take you with us on our journey, please hop on the train, and let's see how we build a digital twin. The basis of our digital twin is an Azure database. All the data of the tunnels are collected in this one.

At first, we started to make a tunnel decomposition. All elements of the tunnel and the geometry, as Jurjen just showed, are put into the system. And then we made failure modes in effect analysis where we look to what might be wrong, what had to be extra attention. And from that, we made a monitoring plan and planned also some research to keep it up to date and up and running.

And after that, we made maintenance plans and also thought of solutions. And these four points are all taken together into the system. So happy using it for us, for our clients, and also for our subcontractors. And what did we do with the digital twin?

The aim was data driven. How to work data driven. The first thing we were thinking about the process was erasing steps from the process we did in an old fashioned way not digital. These steps were necessary in a manual process but not when we were using our digital twin.

Now, you see that the photo in the top left, so all the installations and all the parts had a QR code now, and when we can see it with our handheld. When we do some work or a maintenance engineer is doing the work in the tunnel, he directly put it scanning the QR code and then putting the data into our Azure database, so that will help. The second thing which also bases of the model, when we do some observations you see in the lower left, or sorry, the lower right corner, you see some observations, and in practice, we check it to the design principles and it was not right.

We had the fire extinguishing pipes were eaten by bacteria, so working digital, we could easily see it, put it into the system, and then when we make new designs, we also can prevent that this will occur another time. So that helps. Aim to work that driven. A digital delivery.

When we're talking with our clients, we are often coming into situations where there's a lack of data. It is very important that historic data from design but also historic maintenance data is stored and that we can look back, and from that, do some predictive maintenance in future. And how we do it, we do it with using the ISO 55,000, thinking over how it works and we have from that the same way, because our client also use the ISO 55,000, so we have a similar language.

And we can understand each other how to use and how to work. We also share our thoughts, our risks we see, and what we think is important doing the maintenance to the objects. And that's the basis of the things we do is making it what's critical to quality. We using the lean systematic approach to see what things our clients would be critical to quality.

And that is first listening. Listening very good to our clients. What is the voice of the customer? Making some analysis of it. Analysis. Is he thinking, what would be the worst thing what can occur? If you're installed, it is critical infrastructure.

So the most frightening thing is when the tunnel are immediately has to be closed because it's not safe because there are problems. It is near the vicinity of the airport in Holland. We have a flower auction very connected to the airport, and when you have to drive another road to the airport, it might be too late, and then it might be a problem for the people living around. So from that, we thought of what's critical to quality and doing it in a systematic way.

This was what we did. Our clients, the province had some critical things mentioned in their ISO 55,000, the cost and the traffic flow and all the things mentioned in the blue boxes. Therefrom, we find out what our KPIs were and we gave here an example for costs, and then you can see, what are the critical to quality for tunnel lighting?

And we mentioned some things here, and we are going to use these critical to quality to make a translation from that into our process. And then developing the process and developing the digital twin. You can see it here in the top picture, the current state.

What is the current state? That was how we do it manually, our maintenance for the tunnel. And you see a lot of state, a lot of steps doing some maintenance for the tunnels. And all the rows with red boxes and the current state that are all the steps and the yellow papers that are possible optimizations of it. And from that, we make a new made an evaluation and then we're designing the digital twin.

And you see that at the lower part of the picture, the future state. In this, you see less steps. Some steps totally disappear, and the value added time was increased. And so it was much more effective, because we only have to touch the ball once. Everybody was working with the central data, and there was no rework anymore.

So optimizations we made with the maintenance process was huge. Let's try-- and I will tell you something about how we did it for the tunnel lighting. That was one of our first steps we made in our digital twin. And we started with the architecture of the data model.

We had some internet of things. Automatic information which came within [INAUDIBLE] system and the internet of things at the left side going to the Arcadis framework. And also, field data and asset management to put in it and that made together our main platform. From the Arcadis object type library, we put it all in. The requirements we had for the design and what would be fulfilled with the assets.

And using the Autodesk Construction Cloud in the Autodesk platform, it was possible to make it visible and to analyze it. And at the right side, you see everybody's happy. Our client is happy, our partners are happy, and we are happy because we are able to see and to look at it.

And here, you see lots of data sources, which all has some influence on the tunnel lighting, and it's more than you think, and we are aware that-- so if we make more relations and more analysis it would be learns more about how to optimize use the maintenance and do it predictive for tunnel lighting. And this is then a picture where we put it on. All the fixtures of the lighting has a QR code, and you see it at the left bottom, so it's a connection with the model.

So if you are scanning with the handheld, the QR code, you will know which fixture you have. And then you see at the right, send all information from that fixture will be available. You can see the burning hours here and the number of switches. And from that, it is calculated when one of the lights fails, then we do with Bayesian statistics. We have a apriori distribution from our client and calculating it when there are failures from the lighting, then it's possible to calculate for the group of lighting what will be the rest lifespan there is.

And that helps us doing it most precise to give some thoughts about how could we optimize predictive maintenance. This is a short explanation from it. Here, you will see it. This is our platform, and we are working on the Waterwolftunnel, so we chose that one. And then we chose the lighting groups, so it will zoom in and give all the lighting groups and all the fixtures are within the system.

And then we choose one of the fixtures. This is one, and all the data of that fixture is here in our database. And it's actual at every time you look at it, because if we work on it, the QR code is scanned, and we know it.

So here, we can ask put in a task. So if one of our inspectors sees something and has some questions or a maintenance engineer is doing work on it or replacing a lamp or something, then it will be into the system directly. So at every moment, mostly we work at night. And so when we worked at night, we have the information directly into our system.

So if our office personnel is the next day on the tunnel, we directly know what is done at night, what is the status, and is it-- how does it work and is there enough light in the tunnel to make safe paths possible. And we also did this for tunnel ventilation. Tunnel ventilation is also very important in tunnels, because it's a critical installation, and it's needed for tunnel safety, because if there's a fire, then this big ventilations will be used just to make a maximum flow into the tunnel so everybody behind the fire will be safe.

And here, you see it also in short. We have also QR codes on the ventilation. We have in digital model, so we know where the ventilator is. And so you see some vibrations measurements on all the ventilators, which gives us information if there's a disbalance or something. And we also measured the energy, which is using. So when it's using more energy, there might be something that needs some maintenance.

And all the peaks you see here, that's the starting of the ventilation. So we also know how long does it work and how many hours will be used tunnel ventilation. So this is a picture where the model is working. So we here put our information-- this all into the system using the QR codes.

Someone who is going to look at it will see immediately all the information of the installations, and we also information about the traffic. The cameras are able to analyze digital if we have cars, trucks, or buses, and that's important for us who is using the tunnel. But it's also for example important for the quantitative risk analysis. We have to do by law every year and check if the usage of the tunnel is compliant to what we calculated and expected by law.

And then checking the safety, and now, I will give over to Jurjen. He is going to say this was the model for now, but what will it be for the next 10 years?

JURJEN HAITSMA: Thank you, Bart. Yes. That's the question. We have now several use cases. We know how to exchange data, how to do safe data storage, and we were at the end of the contract.

And that was the nice news last year. The journey goes on. Last September, after we submitted our proposal for AU 23, we won the follow up of the contract. So we have 10 years for new experiences and new investments but also improving quality of life and mobility in the Waterwolftunnel.

A great recognition for our work, the collaboration with the client and the bit we did. Now, we have to the opportunity to maximize our learnings, data, and infrastructure so we can grow and enrich our information model. How can we make the tunnel talk to us, and how are we able to connect the available data, these algorithms, to do a well-predicted maintenance?

So the next step for us are adding specific items, collecting the right data, and identities. Intensively data analytics and artificial intelligence to learn about relations and impacts. Our approach is the Pareto analyzer. First things first, so what are the key cost drivers and what is critical to get to quality? Those are items for our roadmap.

The three to move on our roadmap to a talking tunnel, we are going to add next year three data points, energy supply, fire detection, and fire resistance. With those critical data points, we fuel the infinitely to better understanding of assets and better human experience. We learned a lot, and we shall learn more. And what we learned are that you should really invest in client communication to really understand the client needs, the information he has available, but also the information he needs for his own business processes.

We also learned a lot about data exchange connectivity and security. In first instance, our client talks to us with email and PDFs. Now, we are going to transmit data from our system into their systems. How do you do that in a secure way? How do you make sure that your data is compliant, actual, and accurate?

Then we talk about the benefits. The benefits for the people who operate the tunnel. The real-time tunnel data became real-time tunnel status. So the data became information. Another thing we learned is that we are able with our data create real good business case.

Business case to see how we can reduce our energy consumption and to also provide the right refurbishments on the right time. We hope we have inspired you with our case study and that we have given you ideas to grow your talking asset. And we hope that we will receive the moment that we can ask the tunnel, how are you, and not only in one way, but in different ways. Thank you for your attention.